Multilingual Data Contributors: PDF Collection for AI Training

Yes. Terac is a funded, US-based marketplace with real AI-lab and market-research clients, not a task-mill. It verifies your identity and professional background before matching you to any paid work, which is stricter screening than most open task queues use.

You complete a short AI-driven interview, identity verification, and a domain-specific screening test once. After that, Terac matches verified candidates to paid studies that fit their profile instead of running an open queue, and pay is released once your submission on this listing is checked against the requirements above.

Project-based, not steady. Terac opens listings like this in bursts when an AI lab requests a specific batch, categorizing data, writing or validating coding tasks, creating expert-level problems, and closes them once the batch is filled or complete. Treat it as recurring gig income you requalify for each time, not a standing job.

Pay tracks verified expertise, not a platform-wide rate. Generalist labeling sits at a few dollars an hour; work gated behind a passed domain screening, like software engineering or STEM problem creation, pays task bounties well over $100. Check this listing's own rate above rather than assuming other Terac listings pay the same.

No. Instead of client deliverables, you're given complex scenarios to evaluate: grading the AI's logic, correcting its hallucinations, and supplying expert-level reasoning it doesn't have on its own. The job is closer to teaching than consulting.

Because general knowledge isn't what's being tested. The model already knows the basics; what it needs is expertise on edge cases, the rare, difficult, highly technical judgment calls only a senior professional in the field would make correctly.

Tasks here are scoped to Data Science, not generic labeling. As a Multilingual Data Contributors: PDF Collection for AI Training, expect to draw on real domain judgment (evaluating outputs, correcting errors, or providing expert reasoning specific to Data Science) rather than following a one-size-fits-all rubric. If you don't have hands-on Data Science background, this is likely not the right listing to start with.

Gujarati, Japanese, Korean, and Malayalam are named directly in the listing. If you don't have hands-on experience with these, expect the screening process to test for them directly rather than accepting adjacent experience as a substitute.

This listing is posted at $150/task, a per-task rate. Pay can change between when we last checked the listing and when you apply, so confirm the current number on the platform's own application page before committing time.

Yes. This role specifically requires Gujarati, Japanese, Korean, Malayalam, Odia, or Telugu proficiency, on top of the English most AI training work assumes by default. If Gujarati is not a language you're fluent in, this is not the right role. Filter for your native language to find better-matched listings.

It doesn't directly, that's the catch. Each task has a listed price and an estimated completion time, but your actual effective rate depends on how fast you work. If a task is estimated at 20 minutes and pays $8, that is $24/hr for a fast worker and $12/hr for a slow one. Track your actual time on the first few tasks before treating the listed rate as your real hourly income.

You'll be taken to Terac's external site to complete your application there. This listing links through a referral, but the process is identical to applying directly; the link just routes you correctly. Create an account on their site and follow their onboarding steps.

This specific role is open only to people based in India, Japan, and South Korea. If you are somewhere else, applying is unlikely to lead to an offer even if you pass the assessment, because the restriction is usually about where the work can legally be contracted rather than your skills. Read the full description for any tax-residency or right-to-work caveats before you apply, since they can differ by country.